Corticotropin releasing factor receptor-1 neurons in the lateral amygdala display selective sensitivity to acute and chronic ethanol exposure.

Abstract The lateral amygdala (LA) serves as the point of entry for sensory information within the amygdala complex, a structure that plays a critical role in emotional processes and has been implicated in alcohol use disorders. Within the amygdala, the corticotropin-releasing factor (CRF) system has been shown to mediate some of the effects of both stress and ethanol, but the effects of ethanol on specific CRF1 receptor circuits in the amygdala have not been fully established. We used male CRF1:GFP reporter mice to characterize CRF1-expressing (CRF1+) and nonexpressing (CRF1−) LA neurons and investigate the effects of acute and chronic ethanol exposure on these populations. The CRF1+ population was found to be composed predominantly of glutamatergic projection neurons with a minority subpopulation of interneurons. CRF1+ neurons exhibited a tonic conductance that was insensitive to acute ethanol. CRF1− neurons did not display a basal tonic conductance, but the application of acute ethanol induced a δ GABAA receptor subunit-dependent tonic conductance and enhanced phasic GABA release onto these cells. Chronic ethanol increased CRF1+ neuronal excitability but did not significantly alter phasic or tonic GABA signaling in either CRF1+ or CRF1− cells. Chronic ethanol and withdrawal also did not alter basal extracellular GABA or glutamate transmitter levels in the LA/BLA and did not alter the sensitivity of GABA or glutamate to acute ethanol-induced increases in transmitter release. Together, these results provide the first characterization of the CRF1+ population of LA neurons and suggest mechanisms for differential acute ethanol sensitivity within this region.


Introduction
The amygdala complex has been implicated in a number of important functions, notably emotional processing of internal and external sensory stimuli and the coordination of relevant behavioral output (Pitkänen et al., 1997). Amygdala dysfunction is implicated in anxiety (Tye et al., 2011) and alcohol abuse disorders . The lateral amygdala (LA) serves as the entry point for sensory information and sends excitatory projections to other amygdala nuclei, including the central amygdala (CeA) and basolateral amygdala (BLA), to facilitate stimuli processing (Sah et al., 2003;Agoglia and Herman, 2018). The LA is required for the acquisition and expression of fear learning and memory (Sears et al., 2014), and plays a crucial role in the development of anxiety-like behaviors (Rodrigues et al., 2004). Similar mechanisms may be involved in the dysregulated amygdalar activity seen in alcohol dependence (McCool et al., 2010), but the diversity of cell types within the LA complicates the interpretation of ethanol (EtOH) effects.
GABAergic neurotransmission is sensitive to acute and chronic ethanol exposure, and GABA A receptor activity is involved in ethanol tolerance and dependence. (Eckardt et al., 1998;Grobin et al., 1998;Weiner and Valenzuela, 2006). Both phasic (immediate, short-term inhibition) and tonic (persistent inhibition) GABAergic transmission within the CeA is sensitive to acute and chronic ethanol in a cell type-specific manner (Herman et al., 2013. The functional characteristics of GABA A receptors are determined by their subunit composition; receptors containing the a4, a6, and/or d subunit are expressed extrasynaptically and mediate tonic conductance (Semyanov et al., 2004). These receptors also display an increased sensitivity to ethanol (Wallner et al., 2003;Wei et al., 2004) and may be a primary target for ethanol in the brain (Wallner et al., 2003;Mody et al., 2007), although the direct action of ethanol on tonic GABA A receptors remains controversial (Borghese and Harris, 2007;Baur et al., 2009). Tonic inhibition has been described in principal cells and local interneurons in the LA, but the receptor composition mediating this tonic conductance in LA neurons is unclear (Marowsky et al., 2012).
Corticotropin-releasing factor (CRF) and the CRF receptor-1 (CRF1) are expressed throughout the amygdala (Van Pett et al., 2000;Calakos et al., 2017) and have been implicated in neuroplastic changes related to fear (Hubbard et al., 2007), anxiety (Overstreet et al., 2004;Rainnie et al., 2004), and alcohol exposure (Nie et al., 2004;Roberto et al., 2010;Herman et al., 2013;Lovinger and Roberto, 2013). Notably, activation of CRF1 receptors increases the excitability of BLA neurons to sensory input (Ugolini et al., 2008), and administration of CRF into the BLA increases activation of calcium/calmodulin-dependent protein kinase II (CaMKII)containing projection neurons (Rostkowski et al., 2013). Despite the expression of CRF and CRF1 in the LA and the relevance of the CRF system to the consequences of ethanol exposure, the specific effects of ethanol on the LA CRF1 neuronal population have not been characterized.
Previous work using a transgenic mouse line expressing green fluorescent protein (GFP) under the Crhr1 promoter (Justice et al., 2008) found that CRF1 1 and CRF1 À neurons within the CeA exhibit distinct inhibitory characteristics and differential sensitivity to acute and chronic ethanol (Herman et al., 2013. The CRF1-containing neuronal population within the LA has not been previously characterized, and could be an important determinant of LA activity and output as well as a site of action for drugs of abuse such as ethanol. The current study uses the same CRF1: GFP mice to selectively target and characterize CRF1 neurons in the LA, not to probe the effect of CRF1 activation, which will be the subject of future studies. Here, we combine electrophysiology, immunohistochemistry, and microdialysis to (1) characterize the phenotype of CRF1 1 and CRF1 À neurons of the LA, (2) investigate phasic and tonic inhibitory transmission in LA CRF1 1 and CRF1 À cells, and (3) determine the effects of acute and chronic ethanol exposure on inhibitory control within the LA.

Animals
Experiments were performed in 59 adult (age, 3-6 months; weight, 19-30 g) male transgenic CRF1:GFP mice that express GFP under the Crhr1 promoter, as previously described (Justice et al., 2008). Mice were bred and group housed in a temperature-and humidity-controlled 12 h light/dark facility with ad libitum access to food and water. All experiments were performed in tissue collected from mice between zeitgeber 2 and 7. All procedures were approved by the Scripps Research Institute and the University of North Carolina at Chapel Hill Institutional Animal Care and Use Committees.
After sectioning, slices were incubated in an oxygenated (95% O 2 /5% CO 2 ) artificial CSF (aCSF) solution containing (in mM): NaCl 130, KCl 3.5, NaH2PO4 1.25, MgSO4 1.5, CaCl2 2, NaHCO3 24, and glucose 10 for 30 min at 37°C, followed by 30 min equilibration at room temperature (RT; 21-22°C). Recordings were made with patch pipettes (3-6 MΏ; Warner Instruments) filled with an intracellular solution containing (in mM): KCl 145; EGTA 5; MgCl 2 5; HEPES 10; Na-ATP 2; and Na-GTP 0.2, coupled to a Multiclamp 700B amplifier (Molecular Devices), acquired at 10 kHz, low-pass filtered at 2-5 kHz, digitized at 20 kHz (Digidata 1440A digitizer; Molecular Devices), and stored on a computer using pClamp 10 software (Axon Instruments). Series resistance was typically ,15 MV and was continuously monitored with a hyperpolarizing 10 mV pulse; neurons with series resistance .15 MV or .20% change in resistance during recording were excluded from final analysis. LA neurons containing the CRF1 receptor were identified by GFP expression and differentiated from unlabeled (GFP À ) neurons using fluorescent optics and brief (,2 s) episcopic illumination in slices from CRF1:GFP reporter mice. Electrophysiological properties of cells were determined by pClamp 10 Clampex software online during voltage-clamp recording using a 10 mV pulse delivered after breaking into the cell. Drugs were applied either by bath or Y-tube application for local perfusion. Recordings (V hold = À60 mV) were performed in the presence of the glutamate receptor blockers 6,7-dinitroquinoxaline-2,3-dione (DNQX; 20 mM) and aminophosphonopentanoic acid (AP-5; 50 mM) and the GABA B receptor antagonist CGP55845A (1 mM). All recordings were conducted at room temperature, and all solutions (bath and Y-tube) were prepared and maintained at room temperature.

Immunohistochemistry
Mice (n = 4) were anesthetized with isoflurane and perfused with cold PBS followed by 4% paraformaldehyde (PFA). Brains were dissected and immersion fixed in PFA for 24 h at 4°C, cryoprotected in sterile 30% sucrose in PBS for 24-48 h at 4°C or until brains sank, flash frozen in prechilled isopentane on dry ice, and stored at À80°C. Free-floating 35 mm brain sections were obtained using a cryostat and kept at 4°C in PBS containing 0.01% sodium azide.
Sections were imaged on a Zeiss LSM 780 laser-scanning confocal microscope (10Â objective, tile scanning of LA). All microscope settings were kept the same within experiments during image acquisition. The analyst was blind to the identity of the red fluorescent signal when performing cell counts, and analysis was performed manually in an unbiased manner at four anterior-posterior levels (equidistant sections located À1.00 through À1.70 mm from bregma). Data are presented as the mean 6 SE.

In situ hybridization
Mice (n = 3) were perfused with ice cold PBS/Z-fix (catalog #NC9378601, Thermo Fisher Scientific) after anesthesia with isoflurane. Following perfusion, brains were dissected and immersion fixed for 24 h in Z-fix at 4°C, cryoprotected in 30% sucrose in PBS for 24 h at 4°C, and flash frozen in isopentane on dry ice. Brains were preliminarily stored at À80°C until they were sliced on a cryostat in 20-mm-thick sections, mounted on SuperFrost Plus slides (catalog #1255015, Thermo Fisher Scientific), and stored at À80°C.
Using an RNAscope fluorescent multiplex kit (catalog #320850, ACD), in situ hybridization was performed for Crhr1, Gfp, and Slc17a7. Target retrieval pretreatment as outlined in the manual provided by RNAscope (document #320535, ACD) was performed by first briefly washing prepared slides in PBS. Next slides were submerged in prewarmed target retrieval buffer (catalog #322000, ACD) and kept at a constant temperature between 95°C and 98°C for 10 min. Slides were then removed and immediately rinsed in distilled water twice, and then dehydrated with 100% ethanol. After dehydrations, slices were demarcated with a hydrophobic barrier pen (catalog #310018, ACD) and digested with Protease IV for 20 min at 40°C in a hybridization oven. Next, the RNAscope Fluorescent Multiplex Reagent Kit User Manual (document #320293, ACD) was followed entirely. Last, slides were mounted with Vectashield with DAPI (catalog #NC9029229, Thermo Fisher Scientific). The probes used were Crhr1 (probe target region 207-813; catalog #418011-C2), Slc17a7 (probe target region 464-1415; catalog #416631-C1), eGFP (probe target region 628-1352; catalog #400281), and a negative control (catalog #320751), all from ACD.
Slides were imaged on a Zeiss LSM 780 laser-scanning confocal microscope (40Â oil-immersion, 1024 Â 10 24 ; of LA at approximately À1.46 mm from bregma; 5 mm zstacks). All microscope settings were kept the same within experiments during image acquisition. Background was subtracted from images based on the negative control for each probe, and signal intensity present in DAPIlabeled nuclei after background subtracted denoted positive cells. To perform quantification, ImageJ was used to manually count DAPI-labeled nuclei expressing fluorescently labeled probes in the region of interest (ROI). Next, the percentage of nuclei positive for one or both probes and the percentage of signal colocalization were calculated. The percentage of Crhr1 1 nuclei expressing a marker of interest was determined by dividing the number of colabeled nuclei by the total number of Crhr1 1 nuclei. Quantification was performed on three to four images (approximately À1.46 mm from bregma) from each mouse in an unbiased manner as probe fluorescence was quantified blindly. Brightness/contrast and pixel dilation are the same for all representative images.

Chronic intermittent ethanol vapor inhalation
Mice were placed in ethanol inhalation chambers (La Jolla Alcohol Research) and exposed to chronic intermittent ethanol (CIE) vapor (16 h) followed by air (8 h) daily for 4 consecutive days/week for a period of 4-5 weeks . Before each vapor exposure, CIE mice were injected with a solution of ethanol (1.5 g/kg) and pyrazole (1 mmol/kg, i.p.), an alcohol dehydrogenase inhibitor, to initiate intoxication and maintain constant blood alcohol levels (BALs). Control mice were exposed to room air and received an injection of pyrazole (1 mmol/ kg, i.p.) at the onset of each ethanol vapor exposure. Ethanol drip rate and air flow were adjusted so as to yield BALs averaging 100-250 mg/dl. BALs were measured throughout exposure using an Analox GM7 analyzer. Average BALs for the CIE mice included in electrophysiological recordings were 174.6 6 15.5 mg/dl. Average BALs for CIE mice in microdialysis experiments were 162.7 6 16.5 mg/dl. Terminal BALs were also determined at the time of death when mice were killed immediately after their last ethanol vapor exposure (CIE mice). Another group of mice underwent 3-7 d of withdrawal after their last vapor exposure before being killed (CIE-WD mice).

Microdialysis
Mice (n = 11) were unilaterally implanted with custom fabricated microdialysis probes (0.5 mm regenerated cellulose) aimed at the LA (from bregma: anteroposterior, À1.5 mm; mediolateral, 62.9 mm; dorsoventral, À4.1 mm from dura). However, as some penetrance into BLA is possible, microdialysis results are described throughout as LA/BLA. Mice were perfused with aCSF at 0.2 ml/min and allowed to recover overnight, as previously described (Pavon et al., 2018(Pavon et al., , 2019. The following morning, the flow rate was increased to 0.6 ml/min and allowed to equilibrate for 60 min prior to collection. Dialysate samples were collected at 15 min intervals during a 1.5 h baseline period. Ethanol (1 M) was added to the aCSF perfusate solution for reverse dialysis through the probe, and samples were collected for an additional 1.5 h during the local ethanol exposure period. This dose of ethanol was chosen for consistency with prior experiments using reverse dialysis in rodents, where 1 M was found to induce maximal changes in extracellular GABA and glutamate levels (Roberto et al., 2004a,b).

Statistical analysis
Membrane characteristics and excitability were compared between groups using a two-tailed t test or a oneway ANOVA, where appropriate. Frequency, amplitude, and decay of spontaneous IPSCs (sIPSCs) were analyzed and visually confirmed using a semiautomated thresholdbased mini detection software (Mini Analysis, Synaptosoft). sIPSC characteristics were determined from baseline and experimental drug conditions containing a minimum of 60 events (time period of analysis varied as a product of individual event frequency). All detected events were used for analysis, and superimposed events were eliminated. Tonic conductance was determined using Clampfit 10.2 (Molecular Devices) and a previously described method (Belelli et al., 2009) in which the mean holding current (i.e., the current required to maintain the À60 mV membrane potential) was obtained by a Gaussian fit to an all-points histogram over a 5 s interval. The all-points histogram was constrained to eliminate the contribution of sIPSCs to the holding current. Drug responses were quantified as the difference in holding current between baseline and experimental conditions. Events were analyzed for independent significance using a one-sample t test and compared using a two-tailed t test for independent samples, a paired two-tailed t test for comparisons made within the same recording, and a one-way ANOVA for comparisons made among three or more groups. In the microdialysis experiments, average baseline concentrations of glutamate and GABA were compared in CIE-WD versus AIR controls using twotailed t tests. To examine the effects of acute ethanol administration on LA/BLA dialysate, two-way repeatedmeasures ANOVA (exposure condition Â time) was used to compare air to CIE-WD mice before and after reverse dialysis of ethanol. All statistical analyses were performed using Prism version 5.02 (GraphPad Software). Data are presented as mean 6 SE. In all cases, p , 0.05 was the criterion for statistical significance.

Phenotype of CRF1 1 LA neurons
To validate the fidelity of the CRF1:GFP expression in the LA, we used the RNAscope assay (n = 11 images from three mice) to examine colocalization of Crhr1, the transcript for CRF1, and Gfp, the transcript for green fluorescent protein (Fig. 1A). The number of positive nuclei in the ROI was consistent between groups (Fig. 1B). Approximately 74% of Crhr1 1 neurons coexpress Gfp and 84% of Gfp 1 neurons coexpress Crhr1 (Fig. 1C), indicating substantial penetrance and fidelity, respectively. To identify the phenotype of CRF1 1 neurons in the LA, we performed in situ hybridization in brain sections from CRF1: GFP mice (n = 10 images from three mice) to examine colocalization of Crhr1 and Slc17a7, the transcript for VGluT1. Consistent with GFP expression and the established glutamatergic makeup of the BLA, Crhr1 and Slc17a7 were similarly expressed in the LA (Fig. 1D-F). The number of positive nuclei counted in the ROI was not significantly different between Slc17a7 1 and Crhr1 1 (Fig.  1G). Approximately 60% of Slc17a7 1 neurons coexpress Crhr1, and ;80% of Crhr1 1 neurons coexpress Slc17a7 (Fig. 1H). These data suggest that Crhr1 1 neurons make up a subpopulation of LA glutamatergic cells and that the majority of Crhr1 1 LA neurons are glutamatergic.
The LA is composed of glutamatergic projection neurons as well as local GABAergic interneurons (Sosulina et al., 2006). The results of the in situ experiments indicated that a subpopulation of the CRF1 1 neurons of the LA do not express Slc17a7, suggesting that these neurons are not glutamatergic but may express calcium binding proteins (CBPs) associated with GABAergic interneurons. Work by Calakos et al. (2017) reported that the majority of PV-containing neurons in the BLA also expressed CRF1, but the expression of CBPs in CRF1 1 neurons of the LA is unknown. We examined PV and GFP colocalization in the LA of CRF1:GFP mice (n = 16 sections from four mice) as well as calbindin (CB) and calretinin (CR). For the purpose of clarity, we refer to GFP 1 and GFP À neurons throughout as CRF1 1 and CRF1 À , respectively. We observed expression of CB, CR, and PV interspersed with GFP in the LA ( Fig. 2A-C), but there were more CRF1 1 cells than CBPcontaining cells (Fig. 2D). Consistent with Calakos et al. (2017), we observed that a substantial percentage (;80%) of CBP 1 cells also expressed GFP (Fig. 2E), suggesting that the majority of LA neurons that express CBPs Figure 2. Calcium binding protein expression in CRF1 1 and CRF1 À lateral amygdala neurons. A, Photomicrograph (10Â) of GFP expression (green fluorescence, left), calbindin expression (red fluorescence, center), and merge (right). Scale bar, 100 mm. B, Photomicrograph (10Â) of GFP expression (green fluorescence, left), calretinin expression (red fluorescence, center), and merge (right). Scale bar, 100 mm. C, Photomicrograph (10Â) of GFP expression (green fluorescence, left), parvalbumin expression (red fluorescence, middle), and merge (right). Scale bar, 100 mm. D, Summary of total cells expressing CRF1 (GFP) and CBPs, n = 16 sections from 4 mice. E, Percentage of CBP 1 cells that coexpress CRF1. F, Percentage of CRF1 1 cells that coexpress CBPs. also contain CRF1. However, the percentage of CRF1 1 neurons that contain CBPs was much lower (,20%; Fig.  2F), suggesting that the majority of CRF1 1 neurons are likely not interneurons that express these calcium binding proteins. Together, the results of the in situ and immunohistochemistry experiments identify the CRF1 1 neurons of the LA as a mostly (;80%) glutamatergic population with a smaller (;20%) population of neurons that express CBPs (potentially GABAergic interneurons).

Expression of GABA A receptor subunits
The phasic and tonic conductance of GABA A receptors is dependent on specific subunit configurations and/or expression. We performed double-label immunohistochemical studies examining a1 and d GABA A receptor subunit expression in CRF1 1 and CRF1 À neurons in the LA (n = 12 sections from four mice). The LA contains a significant number of CRF1 1 cells, in contrast with sparse GFP expression in the BLA (Fig. 5A,D). The a1 GABA A receptor subunit has dense expression in the LA (Fig. 5B) and displays colocalization with GFP (Fig. 5C), indicating expression in the majority of CRF1 1 neurons. In contrast, d GABA A receptor subunit expression was greater in the body of the BLA than in the LA (Fig. 5E) and displays minimal colocalization with GFP ( Fig. 5F), indicating little to no expression in CRF1 1 neurons in the LA.

Chronic intermittent ethanol exposure
To examine the sensitivity of LA neurons to chronic ethanol exposure, we subjected CRF1:GFP mice to CIE vapor exposure (4-5 weeks) and CIE followed by 3-7 d of withdrawal (CIE-WD). There were no significant changes in membrane properties in CRF1 1 neurons following ethanol vapor exposure or withdrawal (Fig. 8A), and, consistent with naive neurons, the majority of CRF1 1 neurons from AIR, CIE, and CIE-WD mice exhibited spike accommodation (Fig. 8B). Rheobase was reduced in CRF1 1 neurons from CIE mice (74.74 6 8.63 pA, n = 19 cells from 11 mice) compared with neurons from AIR mice (46.15 6 5.72 pA, n = 19 cells from 6 mice; t (30) = 2.49, p = 0.0187 by unpaired t test; effect size, À28.58 6 11.49 pA; 95% confidence interval, À52.05 to À5.113; Fig. 8C, left). Rheobase did not significantly differ between CRF1 1 neurons from Figure 7. Effects of acute ethanol exposure on phasic and tonic inhibitory transmission in CRF1 1 and CRF1 À lateral amygdala neurons. A, Representative voltage-clamp recording (top) and cumulative probability histogram of interevent interval (bottom) of a CRF1 1 cell during superfusion of EtOH (44 mM). B, Representative voltage-clamp recording (top) and cumulative probability histogram of interevent interval (bottom) of a CRF1 À cell during superfusion of EtOH (44 mM). C, Summary of the change in sIPSC frequency (top) and amplitude (bottom) following ethanol superfusion compared with baseline in CRF1 1 and CRF1 À cells. pp , 0.05 by one-sample t test comparing differences from baseline within cell type; #p , 0.05 by unpaired t test comparing CRF1 1 to CRF1 À cells. D, Representative voltage-clamp recording of a CRF1 1 cell during superfusion of EtOH (44 mM). White dashed line indicates the level of holding current before and after EtOH superfusion. E, Representative voltage-clamp recording of a CRF1 À cell during superfusion of EtOH (44 mM). White dashed line indicates the level of holding current before and after EtOH superfusion. F, Summary of the tonic current induced by ethanol in CRF1 1 and CRF1 À cells. pp , 0.05 by one-sample t test comparing differences from baseline within cell type; #p , 0.05 by unpaired t test comparing CRF1 1 to CRF1 À cells.
CIE-WD mice (64 6 11.85 pA, n = 10 cells from three mice) and neurons from AIR mice (p = 0.4708; Fig. 8C, left). The threshold to fire was also reduced in neurons from CIE mice (À58.58 6 2.35 mV) versus neurons from AIR mice (À49.71 6 1.50 mV; t (30) = 3.34, p = 0.0022 via unpaired t test; effect size, 8.87 6 2.65 mV; 95% confidence intervals, À14.29 to À3.46; Fig. 8C, right) but was not different in neurons from CIE-WD mice (À52.43 6 2.55 mV) versus neurons from AIR mice (p = 0.3353). In addition, we found no differences in action potentials elicited by ascending Figure 8. Effects of chronic ethanol vapor on membrane characteristics and excitability in CRF1 1 and CRF1 À lateral amygdala neurons. A, Summary of membrane characteristics of CRF1 1 LA neurons from AIR, CIE, and CIE-WD mice. B, Relative proportion of CRF1 1 neurons exhibiting spike accommodation from AIR (left), CIE (center), and CIE-WD (right) mice. C, Summary of rheobase at À70 mV (left) and the threshold to fire (right) of CRF1 1 neurons from AIR, CIE, and CIE-WD mice. pp , 0.05 by unpaired t test comparing CRF1 1 neurons from AIR mice to CRF1 1 neurons from CIE mice. D, Summary of action potentials by current injection in CRF1 1 neurons from AIR, CIE, and CIE-WD mice. E, Summary of membrane characteristics of CRF1 À LA neurons from AIR, CIE, and CIE-WD mice. F, Relative proportion of CRF1 À neurons exhibiting spike accommodation from AIR (left), CIE (middle), and CIE-WD (right) mice. G, Summary of rheobase at À70 mV (left) and the threshold to fire (right) of CRF1 À neurons from AIR, CIE, and CIE-WD mice. H, Summary of action potentials by current injection in CRF1 À neurons from AIR, CIE, and CIE-WD mice. Figure 9. Effects of chronic ethanol vapor on phasic and tonic inhibitory transmission in CRF1 1 and CRF1 À lateral amygdala neurons. A, Representative voltage-clamp recordings of CRF1 1 neurons from AIR (left), CIE (center), and CIE-WD (right) mice. B, Summary of sIPSC frequency (left), amplitude (middle), and decay (right) in CRF1 1 neurons from AIR, CIE, and CIE-WD mice. C, Summary of sIPSC frequency (left), amplitude (center), and decay (right) in CRF1 À neurons from AIR, CIE, and CIE-WD mice. D, Representative voltage-clamp recording of CRF1 1 cells from AIR (left) and CIE-WD (right) mice during GBZ superfusion (100 mM). current injection among the three exposure conditions (Fig. 8D). Together, these findings indicate increases in the excitability of LA CRF1 1 neurons following CIE exposure that are normalized under withdrawal conditions. There were no significant changes in membrane properties in CRF1 À neurons following ethanol vapor exposure or withdrawal (Fig. 8E), and, consistent with neurons from naive mice, approximately half of CRF1 À neurons from AIR and CIE-WD mice exhibited spike accommodation (Fig. 8F). No changes in rheobase were observed among CRF1 À neurons from AIR mice (84.44 6 12.37 pA, n = 9 cells from four mice), CIE mice (65.00 6 7.32 pA, n = 8 cells from three mice), or CIE-WD mice (76.67 6 26.03 pA, n = 6 cells from three mice; Fig. 8G, left). The threshold to fire was also comparable in CRF1 À neurons from AIR mice (À50.18 6 1.47 mV) versus neurons from CIE mice (À48.00 6 1.62 mV) and CIE-WD mice (À51.55 6 1.76 mV; Fig. 8G, right). No significant differences in the number of action potentials across current injection steps emerged among CRF1 À neurons from AIR, CIE, or CIE-WD mice (Fig. 8H). These findings indicate no changes in the excitability of CRF1 À neurons following AIR, CIE, or CIE-WD exposure.

In vivo microdialysis
To evaluate baseline transmitter levels following chronic ethanol exposure and withdrawal, we performed in vivo microdialysis in CRF1:GFP mice exposed to AIR (n = 4) or CIE-WD (n = 7). Mice were implanted with 0.5 mm microdialysis probes (Fig. 10A) aimed at the LA. However, as some penetrance into BLA is possible, results are described as LA/BLA (Fig. 10B). There were no significant differences detected between AIR and CIE-WD mice in basal GABA levels (9.2 6 2.1 and 11.9 6 1.3 nM; p = 0.28 by unpaired t test; n = 4-7; Fig. 10C). Acute administration of ethanol (1 M) in the perfusate solution produced significant increases in LA/BLA GABA levels in both AIR and CIE-WD mice as assessed by two-way ANOVA of preethanol and postethanol reverse dialysis (exposure condition Â time) with a significant main effect of time (F (11,99) = 5.585, p = 0.0001), but no significant effect of exposure condition or interaction of time and exposure condition (Fig. 10D). There were also no significant differences detected between AIR and CIE-WD mice in basal glutamate levels (1264 6 310.5 and 1061 6 295.8 nM; p = 0.67; n = 4-7; Fig. 10E). Acute administration of ethanol (1 M in the perfusate solution) produced significant increases in LA/ BLA glutamate levels in both AIR and CIE-WD mice as assessed by two-way ANOVA of pre-ethanol and postethanol reverse dialysis (exposure condition Â time) with a significant main effect of time (F (11,99) = 4.747, p = 0.0001), but no significant effect of exposure condition or interaction of time and exposure condition (Fig. 10F). These data suggest that baseline excitatory and inhibitory transmitter levels in the LA/BLA are not significantly altered by chronic ethanol exposure and withdrawal, and that the responsivity of these transmitter systems to ethanol also remains intact following chronic ethanol exposure and withdrawal.
continued White dashed line indicates the level of holding current before and after GBZ superfusion. E, Summary of tonic current revealed by gabazine superfusion in CRF1 1 cells. F, Summary of tonic current revealed by gabazine superfusion in CRF1 À cells.

Discussion
The CRF1 system in the amygdala has been shown to play an important role in the development of ethanol dependence, but the CRF1 À -containing neuronal population specifically within the LA has not been fully characterized. Here, we report that CRF1 1 neurons in the LA are composed of multiple subgroups, including a small percentage of neurons expressing calcium binding proteins and a larger percentage of glutamatergic neurons. CRF1 1 neurons exhibit distinct membrane properties, minor differences in baseline excitability, and possess an ongoing tonic GABA A receptor conductance that CRF1 À neurons lack. Acute ethanol exposure increases the inhibition of CRF1 À neurons, but the inhibitory control of CRF1 1 neurons is insensitive to acute ethanol. CRF1 1 neurons displayed increased excitability following chronic ethanol; however, neither CRF1 1 nor CRF1 À LA cells displayed alterations in phasic or tonic GABAergic synaptic transmission following chronic ethanol exposure or withdrawal, and basal changes in extracellular GABA or glutamate levels were not observed between exposure groups. Collectively, these results suggest that CRF1 À LA neurons are sensitive to acute ethanol but that changes in CRF1 1 neuronal excitability following chronic ethanol are not due to neuroplastic changes in inhibitory control.
Both phasic and tonic GABAergic signaling regulate the activity and output of amygdala neurons. CRF1 1 LA cells exhibit heightened basal phasic GABAergic signaling compared with CRF1 À cells and an ongoing tonic conductance that CRF1 À cells lack. Subunit stoichiometry regulates the ability of GABA A receptors to mediate tonic inhibition, with the d , a5, and « subunits imparting sensitivity of GABA A receptors to low levels of GABA that are thought to underlie tonic conductance (Stell and Mody, 2002;Stell et al., 2003;Glykys and Mody, 2007).The results of the immunohistochemical studies indicate that CRF1 1 cells predominantly express the a1 subunit and exhibit little colocalization with the d subunit, consistent with previous reports (Wiltgen et al., 2009). Consistent with this observation, the tonic conductance seen in this population was insensitive to application of the d -preferring GABA A receptor agonist THIP. The tonic GABA A receptors in CRF1 1 cells of the LA therefore do not contain d subunits but may contain alternative subunit stoichiometry, such as a1b 2g 2 or a5b g2. In the CeA, the tonic conductance exhibited by CRF1 1 neurons was enhanced by the application of the preferential a1 GABA A agonist zolpidem, suggesting a role for a1-containing GABA A receptors in tonic inhibition in that population. A similar mechanism may regulate tonic conductance in LA CRF1 1 neurons. The d subunit was sparsely expressed in unlabeled LA cells, as seen previously (Pirker et al., 2000), and a tonic conductance in CRF1 À neurons was stimulated by acute application of THIP. These findings suggest that CRF1 À cells express d subunit-containing GABA A receptors that are not active under basal conditions but may be stimulated by agonist activity or heightened concentrations of extracellular GABA.
Previous research has assessed the effects of ethanol on inhibitory signaling within the LA/BLA broadly, but the effects of ethanol on GABAergic signaling and within specific CRF1 1 and CRF1 À populations have not been previously assessed. We observed that CRF1 1 cells are relatively insensitive to changes in inhibitory control induced by acute ethanol; focal application failed to elicit a change in either phasic or tonic inhibitory signaling in this population. As CRF1 1 neurons exhibited heightened phasic and tonic GABA A signaling, these results may suggest a ceiling effect that precludes the possibility of GABA mimetics such as ethanol from further increasing sIPSC frequency or reducing holding current. In contrast, CRF1 À cells demonstrated an increased tonic conductance in the presence of ethanol coupled with a significant increase in GABA release onto these cells. These differences in sensitivity to acute ethanol may be related to GABA A subunit expression differences between the two populations. d -Containing GABA A receptors have heightened sensitivity to ethanol (Wallner et al., 2003;Wei et al., 2004), and the d -expressing CRF1 À neurons exhibited increases in tonic inhibitory control in response to ethanol that the d -lacking CRF1 1 cells failed to demonstrate. The insensitivity of CRF1 1 cells to acute ethanol was also observed in the CeA (Herman et al., 2013), suggesting that this population may have similar GABA A receptor compositions in multiple amygdala nuclei.
In contrast to the selective effects of acute ethanol, both phasic and tonic GABA A signaling in LA CRF1 1 and CRF1 À cells were not affected by chronic ethanol exposure or ethanol exposure and withdrawal. The microdialysis experiments showed that chronic ethanol and withdrawal did not produce adaptations in extracellular GABA or glutamate levels, which may explain the insensitivity of tonic conductance in CRF1 À neurons to ethanolinduced adaptations. Chronic ethanol exposure has been shown to increase basal GABA concentration in the CeA (Roberto et al., 2004a), elevating ambient GABA that is thought to drive cell type-specific changes in inhibitory control . Although the GABA A receptor subunits associated with CRF1 1 and CRF1 À neurons in the CeA and LA are similar, the lack of elevated ambient GABA in the LA likely precludes any chronic ethanol-induced plasticity in inhibitory signaling in either CRF1 1 or CRF1 À LA neurons. Together, these findings may suggest that, unlike the CeA, inhibitory control of CRF1 1 neurons in the LA is relatively preserved following chronic ethanol exposure.
Importantly, following chronic ethanol exposure CRF1 1 neurons displayed reductions in the rheobase and threshold to fire, indicating increased excitability of CRF1 1 neurons but not CRF1 À neurons. Thus, although inhibitory signaling in the CRF1 1 population is relatively insensitive to the effects of acute ethanol, it is sensitive to chronic ethanol in multiple amygdala nuclei (the CeA and LA), making the CRF1 1 population an important target for the actions of ethanol broadly within the amygdala. The results of the voltage-clamp experiments suggest that this enhanced excitability in the CRF1 1 population is not driven by alterations in GABAergic signaling, which may indicate that these changes are instead regulated by ethanol-induced alterations in intrinsic excitability within the LA. Plasticity in glutamatergic signaling within the LA/BLA has been reported following chronic ethanol exposure (McCool et al., 2010), and the LA specifically exhibits alterations in molecular markers of glutamate signaling following chronic ethanol exposure in nonhuman primates (Alexander et al., 2018) and reinstatement of alcohol seeking in mice (Salling et al., 2017). Future work to characterize glutamatergic signaling in the CRF1 1 and CRF1 À populations of the LA both under basal conditions and following chronic ethanol exposure would help to clarify the mechanisms underlying these ethanol-induced changes in excitability.
In the chronic vapor exposure experiments, we did not find evidence for increased baseline phasic GABAergic signaling in CRF1 1 versus CRF1 À cells that was observed in our experiments with& naive mice. The baseline sIPSC frequency in CRF1 1 neurons from AIR, CIE, and CIE-WD mice was lower than what was observed in CRF1 1 neurons from naive mice and higher in CRF1 À neurons from AIR, CIE, and CIE-WD mice (Figs. 4A,B, 9B,C), collectively leading to a loss of significant differences between the two cell populations in the chronic ethanol exposure experiments. As our data indicate that the CRF1 1 cell population is composed of a majority of glutamatergic principal neurons and a smaller subpopulation of interneurons, it is possible that differences in cell subpopulations sampled between the two experiments could account for these different baseline characteristics. However, we did observe tonic inhibition in the CRF1 1 population and not the CRF1 À population in slices from both naive and vapor-exposed mice, which suggests that a similar population of cells was sampled in both sets of experiments. The loss of population differences in phasic but not tonic inhibition may be attributable to the stress of repeated injection, as the air-exposed mice were given daily pyrazole injections as a control for the treatment given to the CIE and CIE-WD groups. It is also possible that exposure to the air chamber, which as a novel environment may be a mild stressor, contributed to the differences between naive and AIR mice in these experiments. The relative sensitivity of phasic and tonic inhibitory control in CRF1 1 cells to repeated mild stress is an interesting avenue for future studies to explore.
Together, these findings suggest that, unlike adaptations in inhibitory control exhibited by other amygdala nuclei (notably the CeA), GABAergic signaling within the LA is intact despite chronic ethanol exposure and/or withdrawal. This resistance to ethanol-induced plasticity in inhibitory control within the LA may play a significant role in the development of alcohol dependence and alcohol use disorders. Sensory information, including external drug cues and internal states such as craving and withdrawal, is relayed first to the LA from the cortex and thalamus; glutamatergic projections from the LA then synapse with CeA, BLA and lateral paracapsular neurons. Our findings suggest that, despite chronic ethanol exposure, inhibitory control of LA CRF1 1 neurons (many of which are projection neurons) remains unchanged, allowing these cells to communicate with downstream amygdalar regions unimpeded. This suggests that neuroadaptations developing in the CeA  and BLA (Läck et al., 2007;Diaz et al., 2011) on chronic ethanol exposure result from local, intrinsic changes rather than from changes in extrinsic inputs from the LA. These findings may have relevance to amygdala circuitry in other contexts, such as fear learning, and may provide insights into other diseases involving amygdala dysfunction, including anxiety and depression. These findings also highlight the heterogeneous cell types within the LA and underscore the need for further cell type-specific characterization of amygdala physiology and pathology.